This really only hurts with UI stuff. Other JS I can write just fine; pure Javascript in the browser has gotten pretty powerful, it all but has jQuery built in now, you've got XPath, a templating language (simplistic, and needs a bit of XSS care, but adequate for small needs) and a lot of other useful things all just sitting there, but if you want a big widget setup there's a huge learning curve to get there, even for an expert programmer.
For the novice, "classic web" would still be a very powerful thing to learn that isn't too hard, but it's hard to find those references now underneath the pile of not quite functional tutorials, broken install instructions, entirely different languages creeping in (Typescript), etc. It's a total disaster for someone trying to learn this way.
I have quite a few non-technical colleagues who know just enough Python to Get Shit Done.
My father (a writer) taught himself R for research purposes and is productive with it.
Universal Paperclips has a super-basic JavaScript codebase[0] but was a really popular game.
Thing is, if you search "how to do X in Python" you're going to get a bunch of resources aimed at professionals because they're the ones looking for that stuff most of the time. That doesn't mean the resources are prohibitively complex, they're just for people solving a different problem.
He initially built Universal Paperclips to learn JavaScript. It's a cool thing, a lot of people played it, and it doesn't rely on an incomprehensible ecosystem. It's two HTML files, four scripts, two CSS files, and an image.
You don't _need_ all the stuff people complain about to do cool things with JavaScript. Especially if you're just getting started with the language.
And I actually think the level of capability you're talking about is within closer reach than most people realize. Certainly it's much easier now than it was even 10 years ago. Probably not something that a non-programmer can accomplish on day 1, but I'm not sure there's many crafts that can be learned that quickly.
- script without a main function
- to personal or ad hoc codebase for non-tech workplace
- to packaged code to cope with namespaces + type hints + general best practices
- to code that reads like good code from the major open source project
is long, but the slope is easy. There are no discontinuities. Basically anyone who can use Python to glue together networkx/sklearn etc. can eventually write software tools that are useful and stable enough to distribute.
And this extends all the way to the job market, where you now need 3 years of experience for a skill most people know how to use in about a week tops, and only lack the exposure of that skill used in a corporate setting. Where people will parrot "you need 3 months to not be a liability! (or more)" when those individuals will easily earn themselves back in the same timeframe. For projects where all you really do is create endpoints, process the most basic of information and expose it in certain ways. For reference, most people in STEM majors need to learn more difficult stuff in a shorter timeframe, with less help.
Their argument was that he didn't know this and that and how is it that I deemed him worthy?
I interjected by saying that during these 30min I get to ask questions I don't have nearly the time to check one's skills throughly. And the same is true for them.
The most I can achieve in this timeframe is check whether the person won't be actively harmful to the project by e.g. being a charlatan or just malicious.
This was news to the delivery manager and they didn't approve of that.
I refrained from replying that our rates vs the market don't really put us in a position to be picky.
But I did tell them that I worked with "net negative" people(with examples) and it's really not about the skills at the moment of hiring.
It astounds me that software dev, historically seen as a field with and for "poor communicators", somehow believes its interviewers can do what top psychologists consider almost impossible to do. Same happens across almost every industry, but really, tech? "Cold hard logic" tech somehow became "yeah we can totally psychoanalyze someone in 30 minutes by talking"?
Worse, the lack of structure across the entirety makes it more of a lottery and a guessing game which is more or less sold as one's personal (in)capabilities. It's far more harmful than people make it out to be.
True story: I had an interview a couple of weeks ago. I was interviewed by two people who were younger than me, and relatively less experienced than me.
My CV has a number of links to repos for various side projects and such. Best I could tell, neither of them looked, at least not more than a minute. Not a single "Your ____ repo looks ____ let's discuss a bit..." Instead I got ovetly basic questions, some of which didn't map to the role / opportunity.
Furthermore, I could only find one of them on GitHub and it was routine stuff (e.g., wedding website). Apparently, neither understood I was interviewing them as well.
I'll never get that hour of my life back.
I have a phrase:
"How you hire is who you hire."
It's sadly funny how many outfits still don't get that.
Most people will either have the skills needed or the capability to learn said skills. Weeding out obvious poor cases is easy, the remainder is semi-arbitrary weeding out of true positives in hopes of weeding potential false positives. Much of higher education alone does a good job at pushing people to learn difficult things in a short timeframe, with little incentive beyond "getting that paper".
Also, hiring is a lot like parenting, but the bad aspects magnified at least a few times. As you say, barely anyone knows what to do. No one has experience. Almost everyone claims to be an expert (not verbally, but they do act like it). Non-obvious bad choices often go undiscovered; as above, most picks themselves aren't bad (they just aren't great), or something else is blamed instead. Pride, neuroticism and personal bias play huge parts in the decisions.
Worst of all, no one actually reads the studies showcasing most of what hiring managers do have close to zero correlation to anything. I'm really just waiting for a mass study where two companies do an initial vetting, then company A throws darts at a dartboard and company B does the whole job interview shtick, and compare the results 2 years down the line. And maybe add a company C which actually does what studies suggest.
"How you hire is who you hire."
And I have a new one...
"You're not only the interviewer, you're an interviewee as well."
Understanding these two basic are essential to a decent hiring process. Ignore these and you're just another fool in a sea of fools.
I once interviewed a guy with a nice, short 1-pager résumé—not from lack of experience, just a solid, terse document. Not just out of school. Had three or four tech jobs in his history.
He listed one research paper near the bottom, that he'd co-authored in grad school.
I found the paper, skimmed it, and asked him about it in the interview (a lot of times in interviews I just ask candidates about stuff that it seems like they'd know about, that I don't know about, to satisfy my curiosity—I have no idea whether this is good or bad interviewing technique, but my intuition is having someone teach you about something you don't know, that they do, is probably at least as good as asking them to tell you stuff you already know) and he told me I was the only interviewer who'd ever brought it up at all.
Moi? I'd leave it. Them not looking tells as much as if they do.
Those that see it otherwise? That to me a culture red flag. It's an easy swipe left.
I had people get visibly angry during interviews where I was the candidate when I called them out on giving me a problem to solve, that they obviously had in their project and spent weeks on figuring out, while I was somehow supposed to do that in less than an hour.
Oddly specific interview questions are a red flag to me, because they indicate someone had a strong opinion on what's the "proper" way to solve a given problem.
What stops you from creating the most simple html page and playing with CSS? That counts as programming to me, there's also scratch for people who are completely new and want to learn.
Today you "gotta" install node, yarn, webpack, react. Then you gotta make a package.json and configure webpack. Or replace webpack with something called "crap". At least that's what all the "for beginners" tutorials that everyone recommends says! Nobody ever said I could just use notepad!!!
But that's not exclusive for frontend.
20 years ago you'd get a random server with Perl. Probably free. 10 years ago you'd get a random server with PHP. You would use FTP and notepad to learn the ropes.
Today you must install some advanced editing software with some plugins, use a terminal, install some runtime or compiler, get some packages, which won't work, so Google tells you to use "sudo" so it works, but that's very suspicious. Stack Overflow says shouldn't be "sudoing" by the way, this is dangerous, you should actually install this "version manager" thing here to install the first tool you installed, so maybe uninstall the first tool, although there's no uninstaller. Oh, and you need this "package manager" tool that install the version manager in the first place. Then you gotta learn to deploy. At this point you're already too old for this shit.
Desktop apps? Good luck with that.
OF COURSE there are simpler solutions, like just making a website with HTML as you could before, or Jupiter Notebooks. But few people today will stumble on that by accident.
Now the scope of what a webpage is capable of is infinitaly larger, and so is the complexity required for using all the new possibilities. If you want to use advanced stuff without learning the basics, than of course you're going to have a bad time.
Now there are like 50 ways to import JS code, most of them don't work most places, a lot of the code itself won't work in a browser at all (node-only), and a great deal of it is a pain to get into a browser if you're not fully bought into Node and probably also serving the whole "app" from Node.
It's about as easy to write your own JS in a browser, now, but it's vastly harder to start using code from the overall JS ecosystem.
Yes, I agree. That's exactly what I said in several parts of my message. However I believe you missed the point and didn't notice the sarcastic tone.
... except you can literally make desktop and mobile applications now by "playing with css and html"! ^electron -- shudder!^
Part of the reason for the JS mess (which people have been bitching about for years and years) is how easy it is to dip your toes into. Match that with how easy it is to create a "gui" now with HTML and CSS, stir in an endless bunch of half-baked support libraries built by people that just learned JS last month and you get the perpetual hell of the JS ecosystem.
And now that it has been adopted by startups and traditional companies, it is here to stay for a while.
Half of my development time when dealing with a new JS project is un-fucking some JS dependency that I never wanted that lives deep in the bowels of npm_modules.
However, If you're referring to 2012 as back in the day, visual basic was by far and away completely deprecated by that point, if anything people would've been using Visual Studio community edition to do net programming and maybe visualbasic.net
Example: https://github.com/jwagemann/era5_in_gee#workflow-overview
When I was four I could operate our TV: turn on, choose channel and adjust the volume. One device. One remote control. Bach then four year olds could also operate phones and record players. When I was five me and my friends could operate a tape recorder, copy music, record music, create tapes and physically share with friends and family.
GameMaker, too, has a very stable API for its proprietary scripting language and, though it is made with 2D gamedev in mind, can work just fine for simple apps for personal use.
Defold is worth mentioning as well
If this guy was told to use `require` instead of `import` it would have most likely just worked. The instructions were wrong so he failed.
Even back then though you still needed to RTFM. The first time I wanted to do a bubble menu (the one that follows your cursor, focusing an item of the menu) I was redirected to Flash.
it seems Excel is the last place like this - to make small pseudo apps, but it comes with its own fair share of problems (mostly calculations or Excel as a database, no sharing)
I think VBA used to be the thing, but the one inside MS Office is "technologically" stuck around year 1996
The web stack sort of killed all that. You can start with HTML and get some results pretty fast but you'll very rapidly hit the limits of what can be done there and suddenly be hit with this giant wall of complexity. NodeJS helps because now you can use the same language on backend and frontend, but even getting to the point where you have a frontend and backend requires mastering a heap of tech with appalling usability (like UNIX). And the modern JS ecosystem is sufficiently complicated that you'll hit problems like the ones described in the article.
One way to fix this is to bring the VB6 model back, by making desktop apps great again. It can be done. There are toolkits that make it easy to write standalone cross-platform apps (e.g. JVM languages+JavaFX, Flutter, the new C# stuff, Qt, etc). There's no fundamental reason distributing desktop apps must be difficult, it's just a neglected area of tooling.
And that's a nice segue to my current startup, which is tackling exactly that problem. We've produced a tool that makes distributing desktop apps as easy as distributing web apps is (or easier!). It takes the output of your build system (if any) as inputs, and builds self-updating packages for Windows, macOS and Linux without needing you to actually have any of those platforms. It can also fully sign those packages itself (without needing macOS or Windows), or self-sign them in which case you get a `curl | bash` type install experience and the PowerShell equivalent for Windows. Finally it produces a simple download HTML page that detects the user's OS and gives them a nice big green button to click.
We've talked to one well known tech firm that wants to integrate this into their products and a driving reason is, in fact, educational. They have a lot of new programmers who use their developer tools and who learn with a desktop toolkit because it's easy to get things on the screen, there's one language to learn, it's all cleanly following the principles they're taught in class (like objects) and so on. But then they hit a wall when they want to show their friends and family what they're doing, or graduate to actually getting real users. Smoothing that on-ramp should help a lot.
At the moment the tool is in private beta, but if you want to try it email me (address in my profile). After public launch it'll be free for open source projects, licensed for commercial projects. Thinking about a feature that automatically bundles source and adds a "view source" button to the apps, to help students who haven't yet mastered uploading code to GitHub.
The real problem is that there's no more room for the technological "middle class". We've gotten to the point where you cannot be SLIGHTLY interested in computers, want to make a quick app for yourself, or make something small to share with the world. You either have to have no interest at all (and live in blissful ignorance) or be so into this type of stuff that you might as well just make it a career. In between these two extremes lies a valley of frustration and needless complexity, where everything is always breaking and changing because we found out what was used 2 years ago is actually bad but THIS TIME IS DIFFERENT! because now we found a solution that is totally going to work forever this time. It's exhausting to even think about.
Many people got their start programming with something like messing with simple Visual Basic stuff. I, and I'm sure many others who are reading this, started with a TI-83 and a desire to cheat on my math quiz. I know that I'm going a little off the rails here, and this has become more of a rant than anything meaningful, but I really, really hate how there's quickly becoming no space for that sort of thing - nowhere for people formerly uninterested in technology and programming to get their feet wet, nowhere to make a quick application to accomplish a small task, etc..